Preeclampsia (PE) affects up to five times more women with pre-existing diabetes mellitus (PDM) than women without it. The present study aimed to identify the effect of the DASH diet on PE incidence (primary outcome) and blood pressure, glycated haemoglobin (GH), serum lipids, glutathione peroxidase (GP), C-reactive protein (CRP – secondary outcomes) in pregnant with PDM. This randomised, controlled, single-blind trial studied sixty-eight pregnant women with PDM throughout prenatal care until delivery (18 weeks) at a public maternity hospital, Brazil. The standard diet group (SDG) received a diet containing 45–65 % carbohydrates, 15–20 % protein and 25–30 % lipids. The DASH diet group (DDG) received the adapted DASH diet with a similar macronutrient distribution, but with a higher concentration of fibres, unsaturated fats, calcium, magnesium and potassium as well as lower saturated fat. Student's t, Mann–Whitney U and the Chi-square tests were used to compare outcomes. PE incidence was 22⋅9 % in the SDG and 12⋅1 % in the DDG (P = 0⋅25). GP levels significantly increased in the DDG (intra-group analysis; mean difference = 1588 [CI 181, 2994], P = 0⋅03) and tended to be different from the variation in the SDG (mean difference = −29⋅5 [CI −1305; 1⋅365]; v. DDG: 1588 [CI 181; 2994], P = 0⋅09). GH levels decreased significantly and similarly between groups (SDG: −0⋅61 [CI −0⋅26, −0⋅96], P = 0⋅00) v. DDG: −1⋅1 [CI −0⋅57, −1⋅62], P = 0⋅00). There was no evidence of a difference in PE incidence at the end of the intervention between the two diets. The DASH diet seems to favour PE-related biochemical markers.